Skip to contentSkip to conversation

Verify if register is writeable before a write operation

RFC PATCH In review Han Nandor 2 participants 4 messages Apr 2, 2019

Subsystems (2)
lore
Share & tools
lore

Han Nandor

author
#0 lore raw

cover Verify if register is writeable before a write operation

Description
-----------
This is an RFC because I don't know if this is a bug or a normal use case. It seems that the function _regmap_raw_write_impl from the regmap framework verifies that a register is writable only using
the callback function, ignoring the other two (max allowed register, register ranges)

Note: As a left right look I did check also _regmap_raw_read function, and it seems that is missing this checks completely.
Is this a problem as well?

Device/Subsystems Impacted
-------------------------
This will impact drivers that end up using raw writes (firmware download, ...)

Testing
-------
Test configuration:
- Kernel Version: 4.14.60 (just for clarification, the patch is rebased on master)
- The testing was done using a driver that has NVMEM support and is using regmap_bulk_write method to write data to registers.
- The valid register range is 0x00->0xFF.

1. Configure a nvcell (4 bytes long) in DT which is outside the valid register range.
2. Write data to NVMEM using the nvcell
3. Verify the result:
3.1 Without the patch the data write in registers is successful with
no error displayed.
3.2 With the patch the data write is denied and an error is displayed.

Nandor Han (1):
regmap: verify if register is writeable before writing operations

 drivers/base/regmap/regmap.c | 9 ++++-----
 1 file changed, 4 insertions(+), 5 deletions(-)
Signature · 1 line
2.17.2
author
#1 lore raw

RFC PATCH 1/1 regmap: verify if register is writeable before writing operations

regmap provides a couple of ways to validate the register range used.
a) maxim allowed register, b) writable/readable register tables,
c) callback function that can be provided by the driver to validate a register. regmap framework should verify if registers are writeable before every write operation. However this doesn't seems to happen in every situation.

The method _regmap_raw_write_impl is only using the writeable_reg callback to verify if register is writeable, ignoring the other two. This can lead to undefined behaviour since this allows to write to registers that could be declared un-writeable by using any other option.

Change _regmap_raw_write_impl to use the regmap_writeable method to verify if registers are writable before the write operation.

---
 drivers/base/regmap/regmap.c | 9 ++++-----
 1 file changed, 4 insertions(+), 5 deletions(-)
drivers/base/regmap/regmap.c +4 -5
index 4f822e087def..42d8404bc8cc 100644
@@ -1493,11 +1493,10 @@ static int _regmap_raw_write_impl(struct regmap *map, unsigned int reg,
 	WARN_ON(!map->bus);
 
 	/* Check for unwritable registers before we start */
-	if (map->writeable_reg)
-		for (i = 0; i < val_len / map->format.val_bytes; i++)
-			if (!map->writeable_reg(map->dev,
-					       reg + regmap_get_offset(map, i)))
-				return -EINVAL;
+	for (i = 0; i < val_len / map->format.val_bytes; i++)
+		if (!regmap_writeable(map,
+				     reg + regmap_get_offset(map, i)))
+			return -EINVAL;
 
 	if (!map->cache_bypass && map->format.parse_val) {
 		unsigned int ival;
Signature · 1 line
2.17.2
maintainer
#2 lore raw
In reply to Han Nandor
Quoted 7 lines · Han Nandor

Description
-----------
This is an RFC because I don't know if this is a bug or a normal use case. It seems that the function _regmap_raw_write_impl from the regmap framework verifies that a register is writable only using
the callback function, ignoring the other two (max allowed register, register ranges)

Please don't send cover letters for single patches, if there is anything that needs saying put it in the changelog of the patch or after the --- if it's administrative stuff. This reduces mail volume and ensures that any important information is recorded in the changelog rather than being lost.

author
#3 lore raw
Quoted 9 lines · Mark Brown
Quoted 7 lines · Han Nandor

Description
-----------
This is an RFC because I don't know if this is a bug or a normal use case. It seems that the function _regmap_raw_write_impl from the regmap framework verifies that a register is writable only using
the callback function, ignoring the other two (max allowed register, register ranges)

Quoted 6 lines

Please don't send cover letters for single patches, if there is anything that needs saying put it in the changelog of the patch or after the --- if it's administrative stuff. This reduces mail volume and ensures that any important information is recorded in the changelog rather than being lost.

Copy that. If is OK, we can still discuss about this using this patch and I can send a single patch once the things are clear.